He Quantified 200 Years of Disruption | Kai Wu on Separating Software Survivors from Value Traps
TL;DR
Kai Wu of Sparkline Capital analyzes why software stocks are trading at historic discounts (10% below market for the first time in 20 years) and presents a framework using 200 years of patent data to distinguish between genuine value opportunities and value traps facing AI disruption.
📉 Software's Historic Valuation Low 3 insights
First-ever market discount
Software stocks currently trade at a 10% discount to the S&P 500, flipping from a 20-year historical average of a 32% premium.
COVID bubble to bust
Valuations peaked during the 2021 low-rate environment and have declined continuously through 2024 to reach five-decade lows.
Data spanning two centuries
Research extending back to 1980 confirms current software valuations relative to the market are at all-time lows.
🪤 The Anatomy of Value Traps 3 insights
Disguised cheapness
Value traps display attractive price-to-earnings ratios while earnings are actually heading toward zero due to technological obsolescence.
The fundamental lag
Stock prices predict disruption years before fundamentals deteriorate, as Blockbuster and Borders maintained revenue per share for years after their stocks began collapsing.
Classic cautionary tales
Radio Shack, Borders, Blockbuster, and McClatchy appeared cheap during disruption by Amazon and Netflix but were actually terminal declines.
📊 Quantifying 200 Years of Disruption 3 insights
Patent-based methodology
Analysis of every USPTO patent since 1790 identifies trending technologies that are pervasive across multiple industries rather than isolated to single sectors.
Automated NLP clustering
Natural language processing automatically groups patents into technology waves and maps corporate exposure through earnings calls, filings, and analyst reports.
Compounding disruption waves
Technological threats stack over time, with retail facing successive assaults from e-commerce, digital media, social media, and now AI simultaneously.
🤖 Surviving the AI Era 3 insights
Code is not the moat
Sustainable software companies possess intangible assets beyond code itself, such as data advantages and network effects, that serve as true defenses against AI disruption.
Unprecedented pervasiveness
AI represents a general-purpose technology potentially impacting all economic sectors simultaneously, unlike prior targeted disruptions.
Survivor patterns
Companies in sectors previously exposed to disruption often survived and thrived by leveraging existing intangible assets to adapt rather than being rendered obsolete.
Bottom Line
When evaluating historically cheap software stocks, investors must look beyond valuation multiples to identify durable intangible assets that determine whether a company will survive AI disruption or become a value trap.
More from Excess Returns
View all
The Market Cares About Fundamentals — Just Not Yours | The Weekly Wrap - 5/31/2026
Hosts Jack Forehand and Matt Ziggler analyze clips from quant strategist Adam Parker, arguing that markets do trade on fundamentals—specifically forward-looking expectations for 2030-2031 rather than current PE ratios—while the penalty for missing earnings estimates has become structurally harsher than the reward for beating them.
The AI Trade Has a Problem | Ben Hunt, Brent Kochuba and Aahan Menon on What Could Derail It
Markets are ignoring geopolitical exhaustion and inflation shocks to focus on AI's transformative potential, creating a disconnect where stocks 'crash higher' despite risks while systematic macro investors focus on resilient economic data.
We Asked Robert Hagstrom Why MPT Forgot What Matters in Investing — and What He Does Instead
Robert Hagstrom argues that Modern Portfolio Theory (MPT) fundamentally erred by defining risk as price volatility rather than margin of safety, creating an institutional framework that prioritizes emotionally comfortable investing over the primary objective of making money, while business owners focus on cash flows and intrinsic value.
Cheap Is a Warning, Not a Thesis | Adam Parker on What This Market Is Really Pricing
Adam Parker argues that buying stocks simply because they appear cheap is an arrogant and ineffective strategy, asserting that markets efficiently price distributions of future fundamentals (2030-2031) rather than current data, while cautioning that only 9% of public companies currently generate meaningful AI revenue despite massive capital expenditure.